Hi bot-kun!
What the fuck is this thing. I keep hearing people talk about it like it's the best filesystem in the universe...
Jayden Bailey
Jacob Allen
That would be btrfs you moron.
Joseph Hernandez
I've used ZFS for years now, starting on FreeBSD with (3) 2tb drives, then built a big boy (for you) filer with (12) 4tb drives. On the new filer, I use Oracle Solaris as the OS, because they created and maintain the ZFS file system, so I went straight to the source. The downside is you have to know how to use Solaris, which is a bit different from Linux. You can do a search for ZFS version and read about the different features that are available. First off, I would recommend ZFS if you are concerned about your data AND you can afford the additional hardware. By additional hardware, I would recommend at least 2 hard drives of the same make and model. The reason is that ZFS offers the ability to do RAID in many different forms, but also with this feature, the added bennefit of checksumming your data when written, with the ability to re-build your data if it's found to have been corrupted. You can read more about what causes corruption, and it's somewhat rare and probably wouldn't affect you, but it does happen and is real. With at least 2 drives, you can create a software mirror, where a copy of the data resides on each hard drive. If a file becomes corrupt, or if an entire hard drive fails, you don't loose your important data. If a file is corrupt, a the file can be rebuilt by running a scan against the drive, and can be repaired from the second copy from the second hard drive. If a drive fails, again, you still have a full single copy on the working drive. This is just an example, there are other very cool drive configurations you can run, RAID 5 (single parity), RAID 6 (double parity), you can assign hot spares, you can create two RAID 5s or RAID 6 and stripe data accross them. You can create RAIDs out of USB drives to use as a more reliable backup (still have the data if one USB fails). And you get the supper cool snapshot ability. I configure mine to create a snapshot weekly, and if I lose a file, screw up a file, or just want to go back in time on a file, I can do that easily. There are a lot of cool things you can do with ZFS. Like I mentioned, I like running ZFS on Solaris, but I don't use that as my main computer, it's a dedicated file server. FreeBSD would be my recommendation if you want to use it on your workstation, and after that, any Linux distro of your choice that makes it available. I know Debian and Arch both have support for it. Oracle maintains a great information library online for using the file system. If you just want to try it, you can install the file system on most linux distros, then try it using a couple of USB drives, or even using several image files using dd or fallocate. One last thing, what's nice about ZFS is it's not hard drive contoller dependant. If you have a hardware RAID array, and the drive controller fails, you better be able to find another exact model, or you can't read the data on your hard drives. With ZFS if you computer mobo goes tits up, just unplug your hard drives, throw them in another system, and rediscover them with ZFS and your data is back online.
Aiden Wilson
Sounds like some interesting features, mostly related to doing software RAID. My question is: what is the advantage to doing this over using some other filesystem + Linux's Logical Volume Management, which can handle software RAID as well? Keep in mind that LVM can also do snapshots and whatnot as well.
wiki.archlinux.org
Lucas Butler
I don't know if I can fully explain the pros and cons of each, as I'm not as experienced with LVM. I would say, after reading the link, that ZFS would be more robust, although like I said, I haven't used LVM much. One main benefit, at least to me, is the checksum of your data when written to validate the integrity of the data. There are additional attributes you can assign or tweak on the volumes you create, such as encryption, compression, and check-summing as mentioned above. You can probably do the same thing with LVM (I know encryption is an option) but with ZFS it's all rolled in. While LVM says it handles snapshots, how many can it handle? I know with mine, I have it take a snapshot every week, and I've had it going back 8+ months before I go and clean it up. I know with some filesystems, having that many snapshots would essentially grind the system to a halt, VMWare's builtin snapshots for example. Also with RAID on ZFS, it's fairly easy to replace a failed drive and rebuild the array. I assume you can do that with LVM, but I'd have to search it. Does LVM support hot-spares, because that's another nice insurance policy to have if one of your drives fail, it can start rebuilding right away. I hope that provides some additional thoughts on the subject, all just my two cents. If your worried about the protection of your data, give ZFS a look, as I believe it offers a more robust solution in that area that LVM does.
Angel Baker
With ZFS being a multilayered solution, you've got many opportunities for tighter integration. What happens in LVM if an error is detected in RAID? Does it just flag a failure, or does it try to automatically correct it?
Robert Brown
Dropped, everything after that isn't worth reading.
The meatgrinder of hardware firmware is known to be horrible, made by barely functional autists on cocaine.
"Just trust the hard drives" is one of the dumbest things I will hear this week.
Aaron Mitchell
Later:
Later:
The cognitive dissonance is strong with this one.
This faggot also wrote the equivalent of winzip which makes him think he knows what's best for a filesystem designed to run on mixed and untrustworthy media.
Mason Wilson
Watch out! Looks like we found ourselves another liberalist!